ABSTRACT
This chapter describes a method for analyzing panel data to indicate which of two intercorrelated variables is more likely to have "causal priority" over the other. A Goal of much scientific research is to identify variables which, when they change themselves, influence other variables. Such influences among variables may be called "causal relationships." The pairs of cross-lagged relationships between height and weight were not different, and they were slightly lower than the simultaneous correlations. To explore further the cross-lagged correlation method, we applied it to a set of panel data dealing with economic behavior. In a series of studies conducted by the Survey Research Center's Economic Behavior Program with support from the Ford Foundation, a national sample of approximately 800 urban families was interviewed in November 1954, and again in November 1955. The network has some interesting characteristics from a sociological standpoint. Dominating almost all the other variables as a determinant was actual change in family income.
